A app to intelligently search through COVID-19 Open Research Dataset (CORD-19)link and find similar papers powered with Machine Learning and NLP. All with a sleek UI.
Download the model file cord19-300d.magnitude (644.95 MB) form here and after unzipping keep it in dir resources in the root directory.
Note: You will need to login into Kaggle.
For now we will only use the Metadata File from Semantic Scholar.
A sample_metadata.csv with only 1000 doc is added for app readiness. Please update the file by running the app and clicking Update Data
.
├── LICENSE
├── README.md
├── data
│ ├── sample_metadata.csv
│ ├── metadata.csv (will be created on first upate run from the app)
| ├── metadata_processed.pickle (will be created on first upate run from the app)
├── resources
│ ├── cord19-300d.magnitude (You need to download this file, keep the name same or change in config.cfg)
├── src
│ ├── cord19_app.py
│ ├── config.cfg
│ ├── data_io.py
│ ├── embedding.py
│ ├── tokenizer.py
│ ├── utils.py
│ └── vectorizer.py
... (excluded others)
cd src
streamlit run cord19_app.py
Note: On first run Click on Update Data button on sidebar of the app, it will take around 5-7 min, wait for it. Then you can run query. This is only required first time and next when you want to update new data from the source.
